To date, human neuroimaging studies have offered important insights into amygdala function using facial expressions of emotion as presented stimuli. Our aim is to augment these initial studies by concentrating our efforts on data collection strategies that will provide information about our individual subjects. Specifically, we will use functional magnetic resonance imaging (fHRI) to test the hypothesis that subjective ratings of ambiguously-valenced facial expressions can be used to predict amygdala response to these expressions as well as to more clearly negative faces (e.g., fearful faces). In addition to amygdala responsivity, we will collect additional measures aimed at more comprehensively defining the psychological profile of subjects who rate facial expressions positively vs. negatively. In addition, we will assess the reliability of our measures by assessing both amygdala fMRI response and valence ratings at three separate time points (within-subject). Finally, the project culminates with a collaboration with the Essex Project V in assessing amygdala signal as it relates to our valence rating task, in a group of 72 well characterized adolescent subjects with varying risk for internalizing problems, and thus dsk for psychopathology. Parallel offline behavioral studies that will determine the generalizability of our task. These offline studies will be cdtical in determining the applicability of this task for future field studies. The proposed studies build upon the work of our laboratory and that of others demonstrating that fMRI tasks involving the presentation of facial expressions can provide useful information about normal amygdala function and its aberrance in psychopathology. These studies will provide much needed data concerning a) individual subject differences in these responses b) the generalizability of these data to adolescents and c) the relationship of these measures with others related to vulnerability to psychopathology.